[Bug] Consensus breaks not identified

[GregTheGreek]

From Martin on gitter:

Ah, I see. Well, I think there's a different problem. See e..g. https://github.com/ChainSafe/Anemone/blob/master/solidity/contracts/booleanOperators.sol#L13 , which just does assembly { pop(and(2,3)) }. Even if all clients produce different results for and(2,3), you won't ever get a consensus error, because it won't influence the end result. You just execute it, and throw away the result. If you want to trigger a consensus vuln, you need to do something with it. Either store result via SSTORE, or change the execution flow if the result isn't what you expected
You could e.g. xor all results into a "sponge" memory variable, and then store that